Research On ETC Traffic Characteristics And Lane Configuration Optimization Of Expressway Toll Station

As a basic part of the expressway,the toll station has an extremely pivotal role.However,due to the contradiction between the rapidly increasing traffic volume and the manual semi-automatic charging method,causeing queuing congestion due to insufficient capacity often occurs at toll stations.So far,in order to effectively alleviate congestion at toll stations,the queuing model has been widely used.The application of queuing theory is based on the premise that vehicles arrive to meet a certain distribution.In the existing research on queuing theory that take traffic flow as the research object,almost all of them only assume that the vehicle arrival obeys the Poisson distribution,but in fact the Poisson distribution only meets the situation that the vehicle arrival rate is low.Therefore,verifying the arrival distribution of vehicles is of great significance for the use of queuing theory and other related researchs.The queuing model is an effective way to alleviate congestion at the micro level.At the macro level,the intelligent transportation system has great potential in alleviating traffic congestion.As an important basis for the operation of intelligent transportation systems,traffic flow prediction has extremely important practical significance.With the rapid growth of China ’s motor vehicles and the full popularization of ETC(Electronic Toll Collection),the traffic volume of expressway toll stations and the number of ETC vehicles will increase significantly.It can be seen that it is of great practical significance for Toll station construction and structural adjustment to analyze the impact of lane configuration on traffic flow with a high proportion of ETC vehicles.Based on this,first of all,this article takes ETC online charging data as the research object and uses the K-S(Kolmogorov-Smirnov test)method to verify the vehicle arrival distribution at different time periods under different traffic levels.It shows that the vehicle arrival distribution of the highway toll station conforms to the Poisson distribution,and λ in different time periods is different.Secondly,based on the vehicle arrival distribution verification results and related literature,a traffic prediction model based on Markov chain,Gaussian mixture model,conditional Gaussian distribution,joint Gaussian distribution and EM algorithm(expectation maximization algorithm)is established.By predicting the traffic by time period,it is concluded that the prediction model has higher applicability and good prediction accuracy when l is [30veh/15 min,85veh/15min).Finally,on the basis of explaining the simulation basis and purpose,the structure and function of the toll station simulation model are introduced,and the toll station simulation model is established using VISSIM simulation software.Through simulation experiments,the influence of different traffic volume,different number of lanes,different ETC ratios and different lane types on the vehicle driving is analyzed under the premise that the proportion of ETC vehicles is extremely high.The results show that under the premise that the proportion of ETC vehicles is [90%,95%],when the traffic flow is within the range of(1400veh/h,1600veh/h],the effectiveness of the mixed lane is better than the MTC(Manual Toll Collection system)lane.When the traffic flow is greater than or equal to 1800veh/h,the effectiveness of MTC lanes is better than mixed lanes.Based on this conclusion,this paper gives the lane allocation scheme under different traffic flow when the proportion of ETC vehicles is [90%,95%].